Antimicrobial Resistance and Biofilm Formation of Escherichia coli Isolated from Pig Farms and Surroundings in Bulgaria

Escherichia coli (E. coli) is a ubiquitous microorganism with pathogenic and saprophytic clones. The objective of this study was to evaluate the presence, virulence, antibiotic resistance and biofilm formation of E. coli in three industrial farms in Bulgaria, as well as their adjacent sites related to the utilization of manure (feces, wastewater in a separator, lagoons, means of transport, and soils). The isolation of single bacterial cultures was performed via standard procedures with modifications, and E. coli isolates were identified via matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) and polymerase chain reaction (PCR). The disk diffusion method was used to assess antimicrobial resistance, and PCR was used to detect genes for antibiotic resistance (GAR) (qnr, aac(3), ampC, blaSHV/blaTEM and erm) and virulence genes (stx, stx2all, LT, STa, F4 and eae). The protocol of Stepanović was utilized to measure the biofilm formation of the isolates. A total of 84 isolates from different samples (n = 53) were identified as E. coli. Almost all demonstrated antimicrobial resistance, and most of them demonstrated resistance to multiple antibiotics from different classes. No virulence genes coding the Shiga toxin or enterotoxins or those associated with enteropathogenicity were detected. No GAR from those tested for quinolones, aminoglycosides and macrolides were found. However, all isolates that were resistant to a penicillin-class antibiotic (56) had β-lactamase-producing plasmid genes. All of them had ampC, and 34 of them had blaTEM. A total of 14 isolates formed strongly adherent biofilms. These results in a country where the use of antibiotics for growth promotion and prophylaxis in farms is highly restricted corroborate that the global implemented policy on antibiotics in human medicine and in animal husbandry needs revision.


Introduction
A famous and widely cited report stated that 10 million people will die each year by 2050 due to antimicrobial resistance (AMR), popularly known as antibiotic resistance [1]. However, scientists, who have devoted more studies to predictive modeling, have emphasized that, fortunately, this is not to be expected with the current rate of increasing mortality due to AMR but will occur under a very specific worst scenario if no action is taken [2]. According to our observations, such misunderstandings are to be expected, because there are not many research articles on that topic, and we rely on reports for citing. However, a Mumbai, India). Because this study was part of research that included culturing in the search for other bacterial genera, single colonies, suspected of having Salmonella spp., were isolated according to ISO 65791:2017. Enriched samples were cultured on XLD agar (M031, HiMedia, Mumbai, India).
For positive controls, we used E. coli ATCC 35218 (American Type Cell Culture Collection, Manassas, VA, USA), as well as E. coli O:157 and E. coli 41 (Collection of the Stephan Angeloff Institute of Microbiology). All isolated colonies were morphologically characterized with the automatic HD colony counter Scan 1200 (INTERSCIENCE, Saint-Nom-la-Bretèche, France).

Identification of E. coli via Spectral Methods (Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry) (MALDI-TOF-MS)
All isolated colonies from Endo and XLD agar were identified via MALDI-TOF mass spectrometer (Bruker Daltonics, Billerica, MA, USA). The essence of the technique is the identification of microorganisms by mapping their unique protein pattern. A small amount of an overnight bacterial mass with a density of 10 4 to 10 6 CFU/mL was mixed with 1 µL of a matrix solution-α-cyano-4-hydroxycinnamic acid (HCCA)-and was placed on the corresponding well of the matrix. The mixture thus made was allowed to dry and was later loaded into the apparatus. Mass spectrometry occurred under constant high vacuum values, and each sample was exposed to short pulses of laser rays (acceleration voltage of 20 kV, mass range of 2.6-20 kDa, laser frequency of 60 Hz and pulsed ion extraction delay of 170 ns). With the energy created from the laser ray, ribosomal proteins were ionized. The molecular fingerprints were comparted with a reference database for ID using the MALDI Biotyper software (Bruker Daltonics, Billerica, MA, USA). The strains identified as E. coli were used for further analysis.

Isolation of DNA of the Bacterial Colonies of E. coli
The strains confirmed for E. coli via MALDI-TOF were recultured, and the total DNA was extracted from single colonies with either the GeneMATRIX Tissue & Bacterial DNA Purification Kit (E3551, EURx Ltd., Gdańsk, Poland) or the GenElute Bacterial Genomic DNA Kit (Sigma-Aldrich, St. Louis, MO, USA) or through crude lysate preparation. The lysates were made by dissolving one bacterial colony in 100 µL of lysis buffer of 0.05 M NaOH and 0.125% sodium dodecyl sulfate (final concentrations), and samples were incubated for 17 min at 90 • C. The DNA concentration and purity were determined with NanoDrop Lite (Thermo Fisher Scientific Inc., Waltham, MA, USA).
For PCR amplification, we used the Color perpetual Taq PCR Master Mix (2×) protocol (E2745, EURx Ltd., Gdańsk, Poland) optimized in our laboratory, as follows: 1 cycle of initial denaturation running at 95 • C for 5 min; a total of 30 cycles of denaturation (at 94 • C for 30 s), annealing (depending on the melting temperature of the primer for 30 s) and extension (at 72 • C for 1 min); and 1 cycle of final extension (at 72 • C for 7 min) and cooling (at 4 • C). Where lysates were used, Tween-20 and gelatine were added to the reaction mix to final concentrations of 0.5% and 0.01%, respectively. The PCR products were visualized in 1.5 or 2% agarose gels. For positive controls, we used the following strains: E. coli ATCC 35218 for the detection of E. coli strains (uidA and yccT), E. coli O:157 containing LT and E. coli 41 (Collection of the Stephan Angeloff Institute of Microbiology) for the detection of eae genes. For the other different E. coli, strains from the Collection of the Stephan Angeloff Institute of Microbiology were used. Table 1. List of primers with their sequences and temperature of melting (annealing) (Tm).

Test for Biofilm Formation
We used the protocol of Stepanović et al. [39] with small modifications, as described in Dimitrova et al. [25]. The biofilms were photodocumented with a microscopic configuration Nikon Eclipse-Ci-L (Nikon Instruments Europe BV, Amstelveen, The Netherlands), and the optical density (OD) was measured at 570 nm by using an ELISA reader ELx800 (BioTek Instruments, Winooski, VT, USA). The classification of Christensen et al. (Table 2) was used again to determine the adherence potential [40]. Table 2. Correlation between the optical density of samples and bacterial adherence [40].

Isolation of Single Bacterial Cultures
Selected colonies from CCA, Endo or XLD agar were used for the spectral identification of the bacterial species. Endo agar is recommended for the confirmation of suspected members of the coliform group. E. coli are expected to have a metallic sheen on this agar. As XLD agar is a selective medium for Salmonella spp., no colonies were suspected to be E. coli.

Identification by MALDI-TOF-MS and PCR
A total of 85 colonies from different samples were identified as E. coli via MALDI-TOF-MS. Later, 84 of them were confirmed with PCR. Isolates that were positive for either the uidA or the yccT gene were accepted as E. coli. Some colonies that did not have a metallic sheen on Endo and that were not suspected for E. coli turned out to be this bacterial species (F4.2 and T1.1), and not all colonies that had a metallic sheen on this agar turned out to be this species. Generally, not all strains of a species isolated with a certain selective nutrient medium have all the expected typical morphological characteristics; therefore, this is not a new phenomenon.

Antibiotic Resistance from the Disk Diffusion Method
Although erythromycin is not used for E coli, we tested it, because there is the potential horizontal transfer of erythromycin GAR to other bacterial species. As can be seen from Tables 3-8, all isolates, except six (WV1.7, WV2.1, WV2.2, WV2.6, SV3.2 and SV3.4), had resistance to at least one agent, and many of them had resistance to multiple antibiotics. The resistance varied in wide ranges-from 0% for cephalosporins to 81% for tetracyclines and other agents. E. coli was isolated from all types of samples. Our results show that almost all isolates had resistance to multiple antibiotic agents, in line with the global tendency of increases in AMR, including in farm animals [5,6]. The antibiotic class that was associated with the greatest developed resistance was tetracycline (81%), followed by penicillins (56%). The percentage of chloramphenicol resistance was very high in the Karnobat farm and much lower in the other two, averaging 42.9%. The resistance to aminoglycoside was 39.3%. The resistance to trimethoprim/sulfamethoxazole followed the same pattern as that of the other unsorted agent, chloramphenicol, averaging 27.4%. Fluoroquinolones, on the contrary, showed much lower resistance in the Karnobat farm, in comparison to the others, and the average was 20.2%. Resistance to the class of macrolides was low (6%), and resistance to the class of carbapenems and cefamandole was absent. Tetracyclines I  I  I  I  I  I  I  I  I  I  I  I  I  S  I  I  I I  I  R  I  I  I  R  I  I  R  R   Macrolides  Erythromycin  I  I  I  I  I  I  I  S  I  I  I  I  I  R   Cephalosporins  Cefamandole  S  S  S  S  S  S  S  S  S  S  S  S  S  S   Fluoroquinolones   Nalidixic acid  S  S  S  I  S  S  S  I Carbenicillin  S  I  I  I  I  I  I  I  S  S  S  S  I    Multidrug resistance (MDR), defined as resistance to three or more antimicrobial classes of the panel tested, was found for 25 isolates (29.8%).

Macrolides Erythromycin
Although there was a variation in the patterns between farms, fecal samples were resistant to the greatest number of antibiotics (e.g., tetracyclines, penicillins and streptomycin). Likely, the fecal bacteria were subjected to more selective pressure due to the direct consumption of antibiotics, and/or the environmental factors could play a role in losing GAR in some other environments. The fewest E. coli were isolated from lagoons and soils, and they had resistance to fewer antibiotics. However, some of them still showed consistent patterns, such as resistance to tetracyclines or, in the case of Karnobat, to chloramphenicol in addition.

Detection of Antibiotic Resistance Genes
We sought GAR for a certain antibiotic class only in the isolates that showed resistance to an antibiotic from this class (including to antibiotics not presented in this study). The results ( Figure 1) show that, from the group of GAR in this study, the isolates were positive only for β-lactamase-producing genes. They were ampC and blaTEM. Out of 56 tested isolates, all the samples had the ampC GAR, and there were 34 isolates that were positive for blaTEM. These results corroborate that ESBL and AmpC β-lactamase production are important resistance mechanisms in members of the Enterobacteriaceae family [18]. Microorganisms 2023, 11, x FOR PEER REVIEW 11 of 23

Detection of Virulence Genes
No virulence genes from the panel STEC/VTEC (stx and stx2all), ETEC (LT, STa, and F4) and EPEC (eae) were detected among the isolates.

Test for Biofilm Formation
Strongly adherent E. coli (14 isolates) was found among all types of samples except the transport vehicle ones (Table 9, Table 10, Table 11 and Table 12 ). Moderately adherent E. coli was present among all types of samples. The Karnobat farm had the most strongly adherent isolates (8), whereas the Veliko Turnovo farm had the least adherent ones (2). Apart from that, there was no correlation in concern to the type of sample and the farm.

Detection of Virulence Genes
No virulence genes from the panel STEC/VTEC (stx and stx2all), ETEC (LT, STa, and F4) and EPEC (eae) were detected among the isolates.

Test for Biofilm Formation
Strongly adherent E. coli (14 isolates) was found among all types of samples except the transport vehicle ones (Tables 9-12). Moderately adherent E. coli was present among all types of samples. The Karnobat farm had the most strongly adherent isolates (8), whereas the Veliko Turnovo farm had the least adherent ones (2). Apart from that, there was no correlation in concern to the type of sample and the farm.

Discussion
The use of antibiotics as growth promoters and for prophylaxis in farms is a highly debated issue. However, there is enough evidence not only for the spread of AMR from livestock, such as swine, to humans (pig feces and wastewater are one of the hotspots for the spread and circulation of AMR and GAR) but also for genetic similarity (clonal types) between resistant bacteria in animals and in humans, some of which are given in a comprehensive review by Sirichokchatchawan et al., 2021 [41]. For example, four sequence types of ESBL producing E. coli shared between humans and pigs were found in Thailand [41]. Plasmid (sub)types and, again, ESBL genes such as blaCTX-M-1 were found to be shared between Dutch pigs and pig farmers [42]. Therefore, the general agreement among policy makers and society is that the disadvantage of creating bacterial AMR outweighs the benefits of antibiotics. Therefore, even though subclinical antibiotic concentrations not only promote growth but also reduce animal morbidity and mortality, numerous bans or restrictions for antibiotic feed additives have been adopted throughout the world [41,43]. The dilemma is excellently described in the work by Chattopadhyay, 2014 [43].
The legislation in Bulgaria is strict, and since approximately the year 2000, antibiotics

Discussion
The use of antibiotics as growth promoters and for prophylaxis in farms is a highly debated issue. However, there is enough evidence not only for the spread of AMR from livestock, such as swine, to humans (pig feces and wastewater are one of the hotspots for the spread and circulation of AMR and GAR) but also for genetic similarity (clonal types) between resistant bacteria in animals and in humans, some of which are given in a comprehensive review by Sirichokchatchawan et al., 2021 [41]. For example, four sequence types of ESBL producing E. coli shared between humans and pigs were found in Thailand [41]. Plasmid (sub)types and, again, ESBL genes such as blaCTX-M-1 were found to be shared between Dutch pigs and pig farmers [42]. Therefore, the general agreement among policy makers and society is that the disadvantage of creating bacterial AMR outweighs the benefits of antibiotics. Therefore, even though subclinical antibiotic concentrations not only promote growth but also reduce animal morbidity and mortality, numerous bans or restrictions for antibiotic feed additives have been adopted throughout the world [41,43]. The dilemma is excellently described in the work by Chattopadhyay, 2014 [43].
The legislation in Bulgaria is strict, and since approximately the year 2000, antibiotics

Discussion
The use of antibiotics as growth promoters and for prophylaxis in farms is a highly debated issue. However, there is enough evidence not only for the spread of AMR from livestock, such as swine, to humans (pig feces and wastewater are one of the hotspots for the spread and circulation of AMR and GAR) but also for genetic similarity (clonal types) between resistant bacteria in animals and in humans, some of which are given in a comprehensive review by Sirichokchatchawan et al., 2021 [41]. For example, four sequence types of ESBL producing E. coli shared between humans and pigs were found in Thailand [41]. Plasmid (sub)types and, again, ESBL genes such as blaCTX-M-1 were found to be shared between Dutch pigs and pig farmers [42]. Therefore, the general agreement among policy makers and society is that the disadvantage of creating bacterial AMR outweighs the benefits of antibiotics. Therefore, even though subclinical antibiotic concentrations not only promote growth but also reduce animal morbidity and mortality, numerous bans or restrictions for antibiotic feed additives have been adopted throughout the world [41,43]. The dilemma is excellently described in the work by Chattopadhyay, 2014 [43].
The legislation in Bulgaria is strict, and since approximately the year 2000, antibiotics in animal husbandry have been allowed only as therapeutics under veterinary supervision and with the demand of reporting it to the competent authorities.

Discussion
The use of antibiotics as growth promoters and for prophylaxis in farms is a highly debated issue. However, there is enough evidence not only for the spread of AMR from livestock, such as swine, to humans (pig feces and wastewater are one of the hotspots for the spread and circulation of AMR and GAR) but also for genetic similarity (clonal types) between resistant bacteria in animals and in humans, some of which are given in a comprehensive review by Sirichokchatchawan et al., 2021 [41]. For example, four sequence types of ESBL producing E. coli shared between humans and pigs were found in Thailand [41]. Plasmid (sub)types and, again, ESBL genes such as blaCTX-M-1 were found to be shared between Dutch pigs and pig farmers [42]. Therefore, the general agreement among policy makers and society is that the disadvantage of creating bacterial AMR outweighs the benefits of antibiotics. Therefore, even though subclinical antibiotic concentrations not only promote growth but also reduce animal morbidity and mortality, numerous bans or restrictions for antibiotic feed additives have been adopted throughout the world [41,43]. The dilemma is excellently described in the work by Chattopadhyay, 2014 [43].
The legislation in Bulgaria is strict, and since approximately the year 2000, antibiotics in animal husbandry have been allowed only as therapeutics under veterinary supervision and with the demand of reporting it to the competent authorities. In poultry

Discussion
The use of antibiotics as growth promoters and for prophylaxis in farms is a highly debated issue. However, there is enough evidence not only for the spread of AMR from livestock, such as swine, to humans (pig feces and wastewater are one of the hotspots for the spread and circulation of AMR and GAR) but also for genetic similarity (clonal types) between resistant bacteria in animals and in humans, some of which are given in a comprehensive review by Sirichokchatchawan et al., 2021 [41]. For example, four sequence types of ESBL producing E. coli shared between humans and pigs were found in Thailand [41]. Plasmid (sub)types and, again, ESBL genes such as blaCTX-M-1 were found to be shared between Dutch pigs and pig farmers [42]. Therefore, the general agreement among policy makers and society is that the disadvantage of creating bacterial AMR outweighs the benefits of antibiotics. Therefore, even though subclinical antibiotic concentrations not only promote growth but also reduce animal morbidity and mortality, numerous bans or restrictions for antibiotic feed additives have been adopted throughout the world [41,43]. The dilemma is excellently described in the work by Chattopadhyay, 2014 [43].
The legislation in Bulgaria is strict, and since approximately the year 2000, antibiotics in animal husbandry have been allowed only as therapeutics under veterinary supervision and with the demand of reporting it to the competent authorities. In poultry farms, most antibiotics are banned even for therapy. An exception in swine farms is the

Discussion
The use of antibiotics as growth promoters and for prophylaxis in farms is a highly debated issue. However, there is enough evidence not only for the spread of AMR from livestock, such as swine, to humans (pig feces and wastewater are one of the hotspots for the spread and circulation of AMR and GAR) but also for genetic similarity (clonal types) between resistant bacteria in animals and in humans, some of which are given in a comprehensive review by Sirichokchatchawan et al., 2021 [41]. For example, four sequence types of ESBL producing E. coli shared between humans and pigs were found in Thailand [41]. Plasmid (sub)types and, again, ESBL genes such as blaCTX-M-1 were found to be shared between Dutch pigs and pig farmers [42]. Therefore, the general agreement among policy makers and society is that the disadvantage of creating bacterial AMR outweighs the benefits of antibiotics. Therefore, even though subclinical antibiotic concentrations not only promote growth but also reduce animal morbidity and mortality, numerous bans or restrictions for antibiotic feed additives have been adopted throughout the world [41,43]. The dilemma is excellently described in the work by Chattopadhyay, 2014 [43].
The legislation in Bulgaria is strict, and since approximately the year 2000, antibiotics in animal husbandry have been allowed only as therapeutics under veterinary supervision and with the demand of reporting it to the competent authorities. In poultry farms, most antibiotics are banned even for therapy. An exception in swine farms is the

Discussion
The use of antibiotics as growth promoters and for prophylaxis in farms is a highly debated issue. However, there is enough evidence not only for the spread of AMR from livestock, such as swine, to humans (pig feces and wastewater are one of the hotspots for the spread and circulation of AMR and GAR) but also for genetic similarity (clonal types) between resistant bacteria in animals and in humans, some of which are given in a comprehensive review by Sirichokchatchawan et al., 2021 [41]. For example, four sequence types of ESBL producing E. coli shared between humans and pigs were found in Thailand [41]. Plasmid (sub)types and, again, ESBL genes such as blaCTX-M-1 were found to be shared between Dutch pigs and pig farmers [42]. Therefore, the general agreement among policy makers and society is that the disadvantage of creating bacterial AMR outweighs the benefits of antibiotics. Therefore, even though subclinical antibiotic concentrations not only promote growth but also reduce animal morbidity and mortality, numerous bans or restrictions for antibiotic feed additives have been adopted throughout the world [41,43]. The dilemma is excellently described in the work by Chattopadhyay, 2014 [43].
The legislation in Bulgaria is strict, and since approximately the year 2000, antibiotics in animal husbandry have been allowed only as therapeutics under veterinary supervision and with the demand of reporting it to the competent authorities. In poultry farms, most antibiotics are banned even for therapy. An exception in swine farms is the prophylactic use of colistin against post weaning diarrhea.

Discussion
The use of antibiotics as growth promoters and for prophylaxis in farms is a highly debated issue. However, there is enough evidence not only for the spread of AMR from livestock, such as swine, to humans (pig feces and wastewater are one of the hotspots for the spread and circulation of AMR and GAR) but also for genetic similarity (clonal types) between resistant bacteria in animals and in humans, some of which are given in a comprehensive review by Sirichokchatchawan et al., 2021 [41]. For example, four sequence types of ESBL producing E. coli shared between humans and pigs were found in Thailand [41]. Plasmid (sub)types and, again, ESBL genes such as bla CTX-M-1 were found to be shared between Dutch pigs and pig farmers [42]. Therefore, the general agreement among policy makers and society is that the disadvantage of creating bacterial AMR outweighs the benefits of antibiotics. Therefore, even though subclinical antibiotic concentrations not only promote growth but also reduce animal morbidity and mortality, numerous bans or restrictions for antibiotic feed additives have been adopted throughout the world [41,43]. The dilemma is excellently described in the work by Chattopadhyay, 2014 [43].
The legislation in Bulgaria is strict, and since approximately the year 2000, antibiotics in animal husbandry have been allowed only as therapeutics under veterinary supervision and with the demand of reporting it to the competent authorities. In poultry farms, most antibiotics are banned even for therapy. An exception in swine farms is the prophylactic use of colistin against post weaning diarrhea.
Nevertheless, in 2010-2016, even newborn suckling pigs still carried as part of their normal intestinal microflora coliforms that had resistance to most tested antibiotic classes, especially to tetracycline and ampicillin. Lactating sows and young pigs had high resistance to streptomycin. Moreover, despite legislative bans, if we compare our results with those from previous studies of swine farms in Bulgaria, an increase in AMR is observed. In the period of 2010-2016, resistance toward tetracycline, ampicillin and streptomycin (approximately 70%, 60% and 65%, respectively) doubled in comparison to the period of 2000-2004 [44]. After a transient decrease in 2020 [25], resistance to tetracycline rose even more to 77.8% and to 81% for the drug class as a whole in our study. Resistance to pefloxacin, carbenicillin and chloramphenicol rose slightly, and there was not a clear correlation regarding the other tested agents [25,44,45].
It is interesting that blaTEM was relatively rare in the study of the period of 2010-2016 [44], whereas in this work, blaTEM was a very prevalent gene with 34 positive samples from 56 tested. This marks an increase from our last time tested in 2020 [25]. AmpC was the only other GAR detected by us with all samples positive out of 56 tested. Last time, it had a similar pattern, because the most numerous positive samples were for that gene [25].
Indeed, there is decreased resistance for some agents, but it still remains relatively high. For example, ampicillin values showed growth in 2021 to 75% but have now decreased to 53.6%. Similarly, streptomycin resistance decreased in 2020 (12.5%) [25] and in this study (39.3%) but still remains high. Amoxicillin resistance decreased in comparison with the period of 2012-2020 [25,45], from 75% to 52.4%. There is very low resistance in farm pigs to third-generation cephalosporins [44] and even to the second-generation agent cefamandole in 2020 [25] and in our study.
As a summary for Bulgarian farm swine, there is high resistance of resident and pathogenic strains to tetracyclines, penicillins and aminoglycosides. Therefore, the high prevalence and increase in AMR in farms with highly restricted antibiotic use (and only for therapy) could indicate residual AMR from past times, the overuse of antibiotics for therapy in farms and/or the high circulation of AMR in the environment due to the high use of antibiotics by humans. This is enhanced by travel and transport in our global society, raising the spread of resistant strains.
The rate of antibiotic resistance differs considerably from country to country globally, depending upon the amount of usage. In the European Union (EU), the lowest levels of AMR E. coli isolates have been found in countries where lower antimicrobial usage is practiced, such as in Norway, Sweden and Finland, whereas countries with high levels of use, such as Spain, Portugal and Belgium, have relatively higher levels of AMR E. coli [46]. For instance, a clear spatial pattern was detected for tetracycline resistance, with high resistance levels reported for southern and western European countries and much lower levels reported for northern and eastern countries in 2004-2007 [7].
In 2019, some antimicrobial classes were assigned the highest priority, with critically important antimicrobials for human medicine only being available for food animals through veterinary prescription [41]. These are fluoroquinolones, third-and fourth-generation cephalosporins, carbapenems, macrolides and polymyxins (colistin). An increase in resistance to these antibiotics in E. coli in animals may indicate a general resistance trend of concern among Gram-negative bacteria originating from the animal reservoir. Carbapenems may not be used in food-producing animals in the present, but it is alarming that resistance genes have been found in pigs, chickens and other livestock [47].
The monitoring and reporting of resistance data from indicator organisms (commensal E. coli and enterococci) to the European Food Safety Authority (EFSA) is voluntary. EFSA reports show that, for a timeframe of 2004-2020, resistance to nalidixic acid is, in general, low, as it was in our studies. However, in the past (2004)(2005)(2006)(2007), large variability was observed in the reported ciprofloxacin resistance median levels (4-24%), with the highest occurrence of 74% reported by Estonia in 2007. In Denmark, legal restrictions have been in place for the use of fluoroquinolones in food animals since 2002, and as a consequence, ciprofloxacin resistance decreased from 3% in 2004 to 0% in 2007 [7]. Wider legal restrictions likely led to the low overall resistance to ciprofloxacin now (median approximately 5%) [48]. It is noteworthy that resistance to third-generation cephalosporins is very low, usually below 1%, in European as well as in local Bulgarian pig farms; therefore, resistance to that agent is still not a threat in animal husbandry, unlike in clinical settings [7,48]. Regarding human infections with E. coli, the European Antimicrobial Resistance Surveillance Network (EARS-NET) reported in 2017 that the highest population-weighted mean resistance percentage in the European Union and the European Economic Area for E. coli that causes serious infections was to aminopenicillins (58.7%), followed by fluoroquinolones (25.7%), third-generation cephalosporins (14.9%) and aminoglycosides (11.4%). In 2017, resistance to carbapenems remained rare in E. coli [49].
In 2019-2020, reports of farm swine from 30 countries showed that high or very high resistance to ampicillin, sulfamethoxazole, trimethoprim and tetracycline was the most common resistance trait observed. MDR was observed in 34.2% (versus 29.8% in our study) of commensal E. coli isolates from pigs. A wide variety of resistance patterns were observed, but mostly to tetracycline, ampicillin, sulfamethoxazole and trimethoprim, often in combination with other substances but rarely with quinolones. About half of the porcine MDR isolates (52.3%) were resistant to all these antimicrobials. Meropenem resistance was not detected in any isolate of indicator E. coli, in line with the results from Bulgaria [48].
Complete susceptibility to 14 antimicrobials tested was higher than that in our study without a statistically significant difference between countries. The aminoglycoside gentamicin had low median levels of resistance through 2004-2020, unlike in Bulgaria. The median levels of chloramphenicol resistance for all reporting countries were moderate in pigs, unlike the high resistance in Bulgaria [48].
There were positive trends (decreases in the level of resistance) in several countries that were possibly due to the documented overall decline in sales of antimicrobials. Most notably, tetracycline resistance has decreased in 15 countries and increased in only two. In 13 countries, there were only decreasing trends, notably in the Netherlands for four agents and in Cyprus for three agents [48]. Indeed, countries such as Denmark and the Netherlands, which both have had massive swine production in recent years, have achieved tremendous reductions in antimicrobial usage while sustaining peak production. Comparable results have been accomplished in Belgium, France, Sweden and the United Kingdom [7,46,48,50]. In contrast, in six countries, there were only increasing trends, and there were increasing trends in Belgium (despite the reduction in antibiotic use), Poland and Romania for three antimicrobials [48].
In reference to the genetic profile, unlike our results, presumptive ESBL producers were more common than AmpC-producers, and isolates with a combined phenotype (ESBL + AmpC) were uncommon. The occurrence of presumptive ESBL, AmpC or ESBL + AmpC producers in commensal E. coli was 1.3% in fattening pigs. In pork, the prevalence of presumptive E. coli ESBL and/or AmpC-producers in meat was less variable, ranging from 0% (Finland and the Netherlands) to 24.4% (Portugal) [48].
It is interesting that, in the period of 2004-2007, AMR to commonly used antimicrobials was higher in porcine E. coli than that in isolates from chickens and cattle, and in most cases, the countries that reported a high occurrence of AMR in E. coli from chickens also had a high occurrence of resistance in E. coli from pigs [7]. This was likely due to the fact that the global average annual consumption of antibiotics for swine (172 mg/kg) is greater than that for cattle (45 mg/kg) and chickens (148 mg/kg) [51].
As a summary for the EU, the EFSA Animal Health and Welfare Panel (2021) revealed clinical swine E. coli isolates with a high proportion of resistance to numerous antibiotics, with a prevalence from 63% to 70% (to aminopenicillins, sulfonamides and tetracycline). However, lower rates of resistance to clinically critical antibiotics (fluoroquinolones and third-generation cephalosporins) were detected [50]. Likely, the latter was the first fruit of the recent Regulation (EU) 2019/61 on Veterinary Medicines and Regulation (EU) 2019/4 on Medicated Feed, stating that antibiotics shall not be applied routinely, nor for prophylaxis, unless in exceptional cases. They should only be applied for metaphylaxis (treatment of animals without signs of disease, which are in close contact with animals that do have evidence of infectious disease) when the risk of spreading infection is very high and there are no other options, as Barros et al., 2023, described [18]. Differences in resistance between countries might reflect the dissemination of certain E. coli types within animal populations and/or differences in the consumption of antimicrobials in animals among countries [7,48].
The recommendations that could make the use of antibiotics as a feed additive unnecessary are several, including improved hygiene and vaccines (although their efficacy varies considerably), biosecurity (measures taken to prevent disease introduction, such as monitoring animals or plant materials that enter the property, as well as sources of water and feed). Numerous alternatives to antibiotics exist-prebiotics, probiotics, phytogenic substances, bacteriophages, etc. However, they have their limits for therapy and are mostly used for prophylaxis [18]. The development of alternatives for the clinical management of the infections of livestock appears to be the need of the hour [43]. There is a need to establish national surveillance programs and effective policies, particularly in certain world regions, to curtail the threat of the evolution of resistant isolates in swine or other livestock production [33,52].
It is clear that more unhygienic farms (e.g., in the developing world) demand more antibiotics in their feed. However, antimicrobials as a feed additive reduce morbidity and mortality in most hygienic farms in the developed world [43]. Whether farm animals are exposed to more infectious agents than animals in their natural habitat is a question beyond the scope of this work. However, it is clear that wildlife has access to more open ventilated spaces and disinfecting sun beams. Therefore, our hypothesis is that, as the number of natural conditions that livestock lives in increases, fewer antibiotics are needed for growth promotion.
Biofilm formation by foodborne pathogens is a serious threat to food safety and public health [53]. As a foodborne pathogen, E. coli can adhere to and form biofilms on most materials and under almost all environmental conditions in food production plants [54]. In this context, this is of importance for irrigation installations and meat processing plants, given the fact that E. coli can survive for months on dry surfaces [55]. Viable pathogens in detached biofilms from contact surfaces can lead to cross-contamination. Environmental biofilms are most often composed of multispecies microorganisms, and mixed biofilm formation can enhance the sanitizer tolerance of foodborne pathogens. E. coli is capable of forming biofilms with other bacterial species, and that could enhance its pathogenic clones' survival in the biofilm community. Biofilm formation in commercial meat plants may be a source of product contamination with no identifiable cause [53].
Even after the cleaning and disinfection processes, the biofilm could still persist. For example, in nursery units in a pig facility after an extensive washing protocol that included disinfection and being kept empty for two weeks, for E. coli and fecal coliforms, reductions of 41% and 51% were observed, respectively; however, they were still found on floors, drinking nipples and feeding troughs [55].
Although we did not find pathogenic clones in the present research, the ability to form strongly adherent biofilms for approximately 17% of the isolated E. coli commensals that were resistant to commonly used antimicrobials and that were MDR strains is alarming, because a detached biofilm can lead to the spread of the AMR to the environment and bacteria in humans through horizontal gene transfer.
As future directions for our research, colistin resistance could be studied because of its prophylactic use against post weaning diarrhea.

Conclusions
Although some antimicrobial agents show a lower level of resistance in Bulgarian porcine farms in comparison to European ones (e.g., ciprofloxacin), the higher prevalence of resistance to aminoglycosides in Bulgaria is alarming, and so is the higher level of resistance to tetracyclines, because it is already high in the European Union. Antibiotic stewardship is the effort to improve how antibiotics are prescribed by clinicians and used by patients. In terms of antimicrobial stewardship programs, national action plans in many countries cover both human and animal health sectors [41]. Legal restrictions lead to positive trends, but our study is an example of the relatively high prevalence and increase in AMR in farms with banned antibiotics as feed additives and prophylaxis. This is likely the result of their overuse for therapy in farms and/or the high circulation of AMR in the environment due to the high usage of antibiotics by humans. Antimicrobial utilization should be more correctly structured as a dosage and course of treatment.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.